Ugh. I sent the patches before I checked email. On 11/10, Roland McGrath wrote: > > > Oh... this adds more naming problems. And this way x86's send_sigtrap() > > can't use this helper easily. > > It's true. Another approach would be to genericize send_sigtrap so that > all arch's can use it, and call that from tracehook_report_syscall_exit. > On x86 and powerpc TRAP_TRACE is the right si_code for step.
But syscall_trace_leave() uses TRAP_BRKPT on x86. Should I change 4/5 ? It is not easy to genericize send_sigtrap(). The code is simple, but the changes in arch/ code are painfull. Besides, we need some "fill info" helper for utrace-ptrace. So. What about the 1-5 patches I sent? > > [PATCH 1/x] introduce user_single_step_siginfo() > > > perhaps user_single_step_siginfo() should go into ptrace.c ? > > Yes, I don't think it needs to be inlined. The new version of default helper is really trivial, an it is only inlined if the arch code doesn't care about the right implementation. Perhaps it can stay in ptrace.h? Oleg.
